Quality Control

Published
  • April 10 2008, 1:08pm EDT

There are probably worse jobs than running a fast-growing independent subsidiary under the umbrella of a global corporation with deep pockets. You’ll get no argument from Sanjay Rao of Identity Systems, which became part of Finnish mobile handset giant Nokia due to the acquisition of Identity Systems’ parent Intellisync in 2006. Most of Intellisync was a fit for Nokia’s core business of mobile software, but Rao, who co-headed the acquisition for Nokia, carved out Identity Systems and made it his own. With backgrounds in both finance and software engineering, Rao immediately took over as general manager at the identity matching pure-play, making him the highest-ranking person with direct hands on the company. As the inside and outside face of the organization, Rao has since overseen eight quarters of strong growth, including almost doubled revenue and profitability. Most lately, Rao has spearheaded Identity Systems’ push into the broader data quality market, as he recently explained to DM Review Editorial Director Jim Ericson.

DMR: Your company’s message until recently was about identity matching and screening. Is data quality just the logical next step?

Sanjay Rao: We have been busy for two years addressing a very successful growth market in identity resolution. Now we are broadening and bringing that expertise to the data quality market with a new product [released in March 2008] called Information Quality Server [IQS]. Fundamentally, if you think of the problem identity matching solves, it’s really a data quality problem. The new data quality products and Information Quality Server broaden the portfolio beyond matching to [make us] a credible next-generation data quality player.

DMR: How do you make the leap from identity resolution to data quality?

SR: Identity resolution is just one module in a data quality solution. If identity resolution is complexity by depth, data quality is complexity by breadth. Complexity by depth was to take matching and searching and essentially do it for very large volumes of data, very high-speed real time, all that stuff. We’ve been doing that for 20 years. But there are at least five or six key aspects to a comprehensive suite of functionality for data quality, which include profiling, matching, address verification, standardization and consolidation - all those things put together. Then you need to process high data and transactional volumes, which has always been our strength in the identity resolution space. A third thing is the ability to support [multiple] languages and international character sets. A fourth aspect is the ability to deploy data quality products in real-time applications. The fifth thing is to be able to provide versatility in a deployment where you can trade off speed, performance and scalability all within the context of real time. Our data quality and identity resolution engines have that built in from day one.

DMR: One module in IQS is about data consolidation, or a single consolidated view of each entity. That sounds a bit like master data management (MDM) to me.

SR: That’s a great point. We believe that in 18 months or two years, data quality and what is MDM today are really going to blend, and you may not be able to separate one from the other. We want to be that agent of change where you can tone down the feature set, the quality and functionality with the same platform. Customers today are deploying two separate types of platforms; we believe that’s a short-term industry evolution. What the market is going to call for is closer to a MDM platform and all the aspects for data quality beyond that.

DMR: But you’re not calling Identity Systems a MDM vendor, are you?

SR: I want to be very clear - by today’s definition we are not a MDM vendor, although our products, both data quality and identity resolution, are being used by customers who have deployed custom applications that today could be constituted as master data management. It’s just that when they did this three or four years ago, there was no commercial off-the-shelf software out there to buy. Also, all the leading MDM vendors use our technology for one of the most important aspects, which is identity resolution and the matching problem. This was the genesis that started us scratching our heads. We thought, if this is important and we have the wherewithal to do this, it’s just a matter of us focusing to provide a complete feature set to bring that to bear. We see a continuum of identity resolution transitioning to data quality, transitioning to MDM and all its derivatives: anti-money laundering, better and improved customer relationship software, compliance and fraud detection applications. To us, these are all more customization for a single vertical than a data management problem. We believe the platform needs to be agnostic across industries and types of entities that one is tracking.

DMR: Do you associate yourself with people and the customer data side versus products or locations?

SR: Today that is correct, and we deal mostly with party data. We’ll also be making investments and improving our product to start integrating product data because our customers are asking for that as well. We hope to bring this to market in the shorter term. But we are not driven by how the industry segments the market. We are more interested in listening to our customers. We’re not categorizing so much by whether it’s MDM or data quality. All we care about is whether it adds value for the customer.

DMR: What are customers asking for then?

SR: Customers across all industries are being hammered away at by three fundamental phenomena: globalization, consolidation and the need to increase efficiency. Globalization is leading to vast amounts of data and transactional volume in a multilanguage context. Consolidation is bringing disparate resources at disparate time frames under the same administrative or corporate umbrella. Efficiency means requirements for ROI and the ability to cross sell and up sell similar products across different P&Ls.

DMR: I’d imagine that globalization is adding a few new wrinkles to data quality when it comes to cultures and languages.

SR: It absolutely is. We have probably the best expertise on the planet to address not just Indian names but Chinese, Japanese and Korean. We address more than 65 nationalities and cultural names. A multinational corporation like HP, which does business in over 100 countries, would like to run one CRM system, or as few as possible, and as few MDM systems as possible with one horizontal utility for data quality. We really want to address that market.

DMR: Some interesting algorithms and fuzzy matching must come out of specific geographies.

SR: Yes, and opposed to other companies, we take a cocktail approach to the problem because no one algorithm for fuzzy logic works every time across the globe. We use four or five techniques at minimum, and we have rules based on expertise from the field, which are specialized for certain countries and ethnicities. It’s a really solid and intangible skill to replicate how the human mind operates. When you have 100 million or a billion records, there is no human capable of managing that. I am surrounded by the smartest people I have ever met who make this company successful. Every day it’s a challenge to me, how much I have missed learning from the people around me.

DMR: IQS also addresses unstructured data. Are you seeing any traction there?

SR: Unstructured data is not the biggest driver in customer decisions right now because there is so much structured corporate data that enterprises and government agencies are not even leveraging. We do believe in another two years unstructured will be as important as structured is today. There is a difference between bleeding-edge [ideas] that can kill you versus leading edge which makes you successful at a certain time. We’d like to see it as a requirement, but if you evangelize something and the customer is not listening, the rational salesman should stop talking about it.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access